Changes in Patterns of Luteinizing Hormone Secretion Before and After the First Ovulation in the Postpartum Mare

Transcription

1 BIOLOGY OF REPRODUCTION 33, (1985) Changes in Patterns of Luteinizing Hormone Secretion Before and After the First Ovulation in the Postpartum Mare BARRY P. FiTZGERALD, 3 HELEN I ANSON, SANDRAJ. LEGAN, and ROBERT G. LOY3 Departments of Veterinary Science3 and Physiology and Biophysics University of Kentucky Lexington, Kentucky ABSTRACT To determine whether luteinizing hormone (LH) secretion during the first estrous cycle postpartum is characterized by pulsatile release, circulating LH concentrations were measured in 8 postpartum mares, of which had been treated with 150 mg progesterone and 10 mg estradiol daily for 0 days after foaling to delay ovulation. Blood samples were collected every 15 mm for 8 h on occasions: 3 times during the follicular phase (Days -, 5-7, and 8-il after either foaling or end of steroid treatment), and once during the luteal phase (Days 5-8 after ovulation). Ovulation occurred in mares 13. ± 0.6 days postpartum and in 3 of mares 1.0 ± 1.1 days post-treatment. Before ovulation, low-amplitude LH pulses ( 1 ng/mi) were observed in 3 mares; such LH pulses occurred irregularly (1-/8 h) and were unrelated to mean circulating LB levels, which gradually increased from <1 ng/ml at foaling or end of steroid treatment to maximum levels (1.3 ng/ml) within 8 h after ovulation. In contrast, 1-3 high-amplitude LH pulses (3.7 ± 0.7 ng/ml) were observed in 6 of 7 mares during an 8-h period of the luteal phase. The results suggest that in postpartum mares LH release is pulsatile during the luteal phase of the estrous cycle, whereas before ovulation LH pulses cannot be readily identified. INTRODUCTION In a variety of species, tonic luteinizing hormone (LH) secretion is characterized by brief, rhythmic episodes of secretion referred to as pulses (Yen et al., 197; Gledhill and Follett, 1976; Baird, 1978; Lincoln and Short, 1980; Rahe et al, 1980; Gab, 1981). Furthermore, it has been demonstrated in several species that steroid feedback control of tonic LU secretion throughout the ovulatory cycle is achieved by modulating LH pulse frequency and amplitude (Yen et al., 197; Goodman and Karsch, 1980; Gallo, 1981). In the equine species, although pulsatile LH secretion has been observed in both intact (Evans et al., 1979) and ovariectomized (Fitzgerald et ai., 1983) mares, it Accepted February 5, Received February 13, 198. Supported in part by a grant from the Grayson Foundation, Inc. to R.G.L. The investigation reported in this paper (No ) was done in connection with a project of the Kentucky Agricultural Experiment Station and is published with the approval of the Director. Correspondence and reprint requests. remains to be determined first, whether the fluctuations in mean tonic LH concentrations throughout the estrous cycle reflect changes in LU pulse frequency, amplitude, or both, and second, whether ovarian steroids exert a feedback control on tonic LH secretion by eliciting such changes. In postpartum mares, serum LH concentrations during the first estrous cycle after foaling are similar to those during the estrous cycle of the breeding season (Whitmore et al., 1973; Geschwind et al., 1975; Loy et al., 198). Additionally, 93% of foaling mares ovulate between 6 and 15 days after foaling (Loy, 1980; Loy et al., 198) and the interval to first ovulation can be extended by administration of progesterone and estradiol (Loy et al., 198). During such treatment LH is suppressed, but following cessation of treatment serum LH increases and results in ovulation. Furthermore, the interval to ovulation following cessation of steroid treatment is not different from that in untreated postpartum mares following foaling (Loy et ai., 198). Using both untreated and steroid-treated postpartum mares to manipulate the timing of sample collection periods, the present study was performed to 316 Downloaded from on 19 January 018

2 PULSATILE LH RELEASE IN POSTPARTUM MARES 317 determine whether changes in tonic LH secretion during the first estrous cycle postpartum are associated with corresponding fluctuations in LH pulse frequency, amplitude, or both. MATERIALS AND METHODS Animals and Collection of Blood Samples Eight mares, which foaled between May 6 and July 9, 198, were kept at pasture except during periods of frequent sampling when they were housed in stalls with their foals. All mares received oats and hay daily and water was freely available. The occurrence of the first ovulation postpartum was delayed for 0 days in mares by daily injection of progesterone (150 rng) and estradiol-17j3 (10 mg) in 3 ml of cottonseed oil (Lm.). The first injection was administered within h after foaling. To determine the occurrence of ovulation following foaling or end of steroid treatment, all mares were palpated per rectum at least every days beginning on Day 5 postpartum, and daily following the appearance of a follicle 5 mm in diameter or larger. Estrus was detected by teasing daily with a stallion. Beginning on the day of foaling, blood samples (10 ml) were collected daily for 35-0 days by jugular venipuncture into heparmnized tubes. Additional samples (5 ml) were collected every 15 mm for 8 h from an indwelling jugular venous catheter on three occasions, Days -, 5-7, and 8-il after parturition or cessation of steroid treatment. Furthermore, to determine if the sampling interval influenced the secretory LH profile, samples were collected every 5 mm for h after the first h of sampling at 15-mm intervals on Days 8-il. Thereafter, samples were collected every 15 min for more h. Finally, in those mares that ovulated, as determined by rectal palpation, blood samples (5 ml) were collected every 15 min for 8 h on one occasion between Days 5 and 8 after ovulation. Cannulae were inserted between 30 and 60 min before blood collection and maintained patent by flushing with saline (9 g NaC1/I) containing heparmn (5 IU/ml). At each sample withdrawal, the first 1 ml was discarded and a further 5 ml were collected. All blood samples were centrifuged and plasma from daily samples or serum from frequent sampling periods was collected and stored frozen until assayed. Radio immunoassay of LH Luteinizing hormone concentrations were measured in duplicate 10-l00-il aliquots of serum or plasma using a radioimmunoassay procedure described previously (Loy et al., 198). An equine pituitary preparation, kindly supplied by Dr. H. Papkoff, was used as a standard. This LH has.97 times the activity of NIH-LH-Si according to the ovarian ascorbic acid depletion assay. The sensitivity of the radioimmunoassays was 0.07 ng/ml for 00 il serum. Intra- and interassay coefficients of variation were 8% and 9%, respectively, as determined in six replicate 50-SAl aliquots of an intact mare serum pool that produced a 50% inhibition of binding of radiolabeled LH to antibody (13 assays). Analysis of Data The literature contains numerous methods for defining the occurrence of LB pulses (e.g., Baird, 1978; Gab, 1981). One technique employed for detection of LH pulses in a variety of species distinguishes a change in the secretory pattern of LB from within-assay variation by comparing the mean percent coefficient of variation of all LH values composing the ascending and descending portions of the potential LH pulse with that of the within-assay variation (Gallo, 1981). In the present study, the mean percent withinassay coefficient of variation was calculated for 3 serum pools that contained., 5.1, and 13.6 ng/ml, as determined in 10 assays. This range of LB concentrations is similar to that observed in mares during the estrous cyde. Each serum pool was assayed in 3 duplicate determinations at the beginning, middle, and end in each of 10 assays. The mean percent coefficients of variation were 8., 10., and 8.0, respectively. A similar value was obtained when a serum pool containing 1.8 ng/ml was assayed in 50 replicate determinations in a single assay (within-assay coefficient of variation, 7.0%). Thus, over a wide range of LH values the within-assay coefficient of variation averaged 10% or less. The mean percent coefficient of variation for all LH values determined in the respective 8-h sampling periods was calculated. When this value was at least 1.5-fold greater than the within-assay. variation the LH secretory profile was considered pulsatile. The mean coefficient of variation of LH values composing the ascending and descending portions of each potential LH pulse was similarly calculated and compared to the within-assay variation. The amplitude of an LH pulse was defined as the peak concentration during the episode minus the lowest concentration detected among the three samples before the peak. Statistical comparison of the preovulatory increase in plasma LH concentrations, determined in samples collected daily from steroid-treated and untreated mares, was made by a split-plot analysis of variance with days as subplots (Statistical Analysis Systems, SAS Institute, Inc.). Orthogonal polynomials were employed to determine that this increase followed a quadratic function. RESULTS Daily LH Profiles and Timing of Ovulation In the untreated mares, all of which ovulated, plasma LU concentrations in samples collected daily for 0 days postpartum increased gradually from 0. ± 0. ng/ml at the time of foaling to 7.3 ± 0.5 ng/ml on the day of ovulation, 13. ± 0.6 days after parturition. The maximum mean LH concentration of 1.9 ± 1.0 ng/ml was observed 8 h after ovulation. Representative patterns of LI-I secretion in a nontreated individual mare (Mare 9) and in an individual steroid-treated mare (Mare 58) are depicted in Fig. 1 (upper and lower panels, respectively). During treatment with estradiol Downloaded from on 19 January 018

3 1\Jf\ 318 FITZGERALD ET AL. 0.1 o T,roe (doys) Time (days) FIG. 1. Changes in daily plasma LH concentrations in a nontreated (Mare 9, upper panel) and a steroidtreated (Mare 58, lower panel) foaling mare. Day 0 represents the day of foaling and the batched areas indicate the days of estrus. The day of ovulation is indicated by t, and the arrows labeled A, B, C, and D indicate the days of frequent sampling before and after ovulation. For Mare 58, the stippled bar indicates the duration of steroid treatment (see text). concentration were not consistently observed in the respective sampling periods. Furthermore, the number of LU pulses remained low (1-/8 h), and therefore seemed unrelated to mean circulating LH concentrations, which increased in the 7 mares from. ± 0.5 ng/ml on Days - to.3 ± 0.5 ng/mlon Days 8-11 (P<0.05). Finally, the detection of LH pulses was not influenced by the frequency of sample collections, as illustrated by the absence of LH pulses in mares (Mares 31 and 53) from which samples were collected every 5 mm for a -h period on Days 9 and 11 postpartum, respectively (Fig., right panels). In one of these mares (Mare 53; Fig., lower right panel) the mean LH level was 6. ng/ml; this elevated concentration was associated with ovulation during the sampling period, as determined by rectal palpation of the ovaries before and immediately after the 8-h sample collection period. Luteinizing Hormone Pattern in the Absence of Ovulation and progesterone, plasma LH concentrations remained low (< 1 ng/ml). Following cessation of treatment, only 3 of the mares ovulated. Luteinizing hormone concentrations in these 3 mares increased from 0.5 ± 0.1 ng/ml to 8.8 ± 1. ng/ml on the day of ovulation, 1.0 ± 1.1 days later. Maximum LH concentrations of 11. ± 1.0 ng/ml were attained h after ovulation. The periovulatory increase in plasma LH concentrations in steroid-treated mares was not different (P>0.05) from that in untreated mares; therefore, the data were pooled for comparisons between time periods. When mean LH levels in the 7 mares that ovulated were normalized to the day of ovulation, the preovulatory increase in LH manifested a slow rate of rise from around 1 to ng/ml in 8 days (Day -10 to Day -). Thereafter, mean LU levels increased abruptly from to 13 ng/ml during the next 3 days, with peak levels occurring 1 day after ovulation (Fig. ). During the preovulatory period LH pulses were only identified in 3 mares. Representative examples of the typical pattern of LU release in 3 mares in which LU pulses were not identified are illustrated in Fig. 3. In those mares in which LH pulses were identified (Fig., left panels), such identifiable increases in LH In contrast to the low-level fluctuation in serum LH concentrations observed before ovulation, LH secretion in the mare that failed to ovulate was low and stable. This pattern of LH secretion was particularly evident in samples collected every 5 mm during a -h period 10 days after cessation of steroid treatment (Fig. 5). This absence of any measurable 0 C r6 -j ( 10 0 N 7 I I I P I Oays From Ovulation 0 FIG.. Changes in mean (± SEM) plasma LB concentrations in samples collected daily from 7 mares between 10 days before and days after ovulation. Values were normalized to the day of ovulation. Downloaded from on 19 January 018

4 PULSATILE LB RELEASE IN POSTPARTUM MARES TIME (hours) FIG. 3. Serum LH concentrations in samples collected every 15 min from 3 representative mares that did not exhibit LB pulses before ovulation, illustrating the typical low-amplitude fluctuation. The day of frequent sample collection postpartum or post-treatment is indicated at the upper left of each profile. fluctuation in LU concentration was associated with only minor changes in the daily LU profile, since plasma LU levels increased to a maximum of 1.1 ng/ml 5 days after the end of steroid treatment and decreased to < 0.5 ng/ml within 3 days, remaining at this level for the rest of the study. Additionally, only minor follicular development, as judged by rectal palpation of the ovaries, was observed during the first 5 days after the end of steroid treatment, at which time a follicle approximately 0 mm in diameter was identified. This follicle subsequently regressed by Day 7 and thereafter the ovaries were devoid of any palpable structures. Luteinizing Hormone Release after Ovulation By Days 5-8 after ovulation, plasma LU concentrations in samples collected daily had decreased to. ± 0.9 ng/ml (n=7). During this period frequent collection of samples for 8 h revealed a pulsatile pattern of LU release in 6 mares. Representative examples of the highamplitude LU pulses (3.7 ± 0.7 ng/ml, n=11), which occurred at a frequency of between 1 and 3 pulses per 8 h, are illustrated in Fig. 6. A rapid and pronounced increase in serum LH concentration was associated with each pulse, and peak values were attained within 5 mm of the previous nadir. In the mare that exhibited 3 LU pulses during 8 h (Mare 9; Fig. 6, upper left panel) peaks of the pulses occurred at 165-mm intervals. In the remaining mares the number of LU pulses was too low to estimate the inter-pulse interval, although the suggestion of a third LH pulse in one mare at the end of the sample collection period also revealed a constant interval between successive peaks (10 mm) (data not shown). Finally, in the one mare in which LU pulses were not observed after ovulation, serum LU concentrations gradually decreased throughout Downloaded from on 19 January 018

5 30 FITZGERALD ET AL TIME (hours) FIG.. Serum LH concentrations from representative mares that exhibited LB pulses (v) before ovulation. The day of sample collection postpartum or post-treatment is indicated at the upper left of each profile. In each profile samples were collected at 1 5-mm intervals except between Days 8 and 11 postpartum or post-treatment (right panels), when samples were collected every 5 mm between hours and 6 of sampling. the 8-h period, suggesting that an LU pulse had occurred before initiation of sample collection DQy101 I I (data not shown). In support of this possibility, the LU value at the beginning of the 8-h period of frequent sample collection was higher than E 0 -.t-s--s-- -SSS that observed in a blood sample collected 0 min earlier for the purpose of determining the daily LU profile. Thus, although LH pulses were observed only in 6 mares, it is likely that LU C -J 0 - OQy 3 pulses occurred in all 7 mares after ovulation. DISCUSSION a -I I I I I - The observed changes in the mean patterns of LU secretion postpartum and following cessation of steroid treatment, as revealed in samples collected daily, compare favorably with those reported previously (Nett et al., 1975; Loy et al., 198). The present report extends these studies by attempting to determine whether the fluctuations in mean tonic LH concentrations before and after ovulation in the TIME (hours FIG. 5. Serum LH concentrations from one mare (Mare 7) that did not ovulate after cessation of steroid treatment (see text), illustrating low LH levels that were associated with only minor follicular development during the experimental period. The day of sample collection post-treatment is indicated at the upper left of each profile. Downloaded from on 19 January 018

6 PULSATILE LB RELEASE IN POSTPARTUM MARES TI ME ( hours) FIG. 6. Serum LH concentrations in representative mares during the luteal phase of the first estrous cycle postpartum, illustrating the occurrence of LH pulses (v). Samples were collected at 15-mm intervals. postpartum mare reflect changes in LU pulse frequency, amplitude, or both. During the preovulatory period of several species including cows, sheep, and humans, LU is secreted in discrete, high-frequency pulses that are lower in amplitude than luteal phase pulses (Yen etal., 197; Rahe etal., 1980; Baird and McNeill, 1981). In the ewe, it has been demonstrated that the frequency of blood collection during the preovulatory period is an important factor since blood samples collected every 1 mm underestimate the frequency of LU pulses compared to that determined in samples collected every mm (Karsch et al., 1983). Thus, in the ewe, samples collected at 1-mm intervals during the preovulatory period reveal a sawtooth pattern of LU secretion, and the individual LH rises that contribute to this pattern of LU secretion can be shown to constitute a portion of discrete low-amplitude pulses, indicating that this sampling frequency used in this study did not enhance the ability to detect secretory pulses. the postpartum mare was not pulsatile. In further contrast to the ewe, samples collected from preovulatory mares at 5-mm intervals did not increase the number of identifiable LU pulses, indicating that this sampling frequency used in this study did not enhance the ability to detect secretory pulses. In contrast to the lack of detection of LU pulses during the preovulatory period, when mean LU levels are increasing, we have shown previously that in ovariectomized mares changes in LU pulse frequency are an important determinant of circulating LU concentrations (Fitzgerald et al., 1983). During the transition from anestrus to the breeding season, mean serum Downloaded from on 19 January 018

7 3 FITZGERALD ET AL. LU levels in ovariectomized mares increase approximately 9-fold (from 1 to 9 ng/ml) in association with a -fold rise in LU pulse frequency. However, during the breeding season, when LH levels are maximal, LU pulses can no longer be identified, and at this time LU secretion is characterized by rapid, low-amplitude fluctuation. Because this variable pattern of LH release is preceded by a period when LU pulses, the frequency of which gradually increases, can be readily identified, we have proposed that the elevated LH levels during the breeding season reflect high-frequency LU pulses. Thus, it may be postulated that, in intact mares during the breeding season, whenever ovarian negative feedback is either reduced or absent, such as during the preovulatory period (Ginther, 1979), low-amplitude, highfrequency pulses of LU release might also occur. In contrast, the present results demonstrate that, during the preovulatory period in the postpartum mare, LH release is not pulsatile, at least as determined in blood samples collected from the jugular vein. If so, such a mechanism for the control of tonic LU levels contrasts markedly with that in other species in which changes in circulating LH concentrations are governed by alterations in pulse frequency and amplitude (Yen et al., 197; Goodman and Karsch, 1980; Gallo, 1981). One explanation to account for the absence of identifiable LU pulses during the preovulatory period of the postpartum mare might be that the long half-life of equine LU (about 1 h; Ginther et al., 197) masks low-amplitude LU pulses. Alternatively, LH pulses may occur in postpartum mares during the preovulatory period, but a different pleiomorphic form of LH is secreted than that in ovariectomized and luteal-phase mares. Such pulses might not be detected by our assay. However, on the basis of preliminary observations of an increase in frequency and decrease in amplitude of LU pulses in intact mares entering the breeding season (Fitzgerald, unpublished), it is most likely that the low-amplitude LU fluctuations observed during the preovulatory period are indicative of high-frequency, low-amplitude LH pulses. Thus, in certain physiologic states the frequent collection of blood samples from the jugular vein might not provide an accurate indication of LU release by the anterior pituitary gland. The necessity for further studies on the mechanisms governing LH release in the mare is further emphasized by the observation of discrete high-amplitude LU pulses during the luteal phase of the first estrous cycle postpartum. One reason to account for this striking change in the pattern of LU secretion between the pre- and postovulatory periods is the changing steroid feedback environment. In this regard, progesterone is known to exert a negative feedback effect on tonic LU release in the mare, and the action of this steroid is enhanced by estradiol (Garcia et al., 1979). The present study demonstrates that LU pulses occur during the luteal phase of the estrous cycle when progesterone is elevated. Thus, in the mare as in other species, progesterone may exert a negative effect on tonic LU secretion by modulation of LH pulse frequency and/or amplitude (Goodman and Karsch, 1980). However, this proposal is based upon the assumption that the low-amplitude fluctuation in tonic LU levels before ovulation, when progesterone is absent, is due to high-frequency, low-amplitude LH pulses. If LH is secreted in a nonpulsatile fashion before ovulation, as discussed earlier, then the secretion of progesterone may be considered as an inducer of pulsatile LU release rather than as acting by modulating LU pulse frequency. In conclusion, the present results suggest that in the postpartum mare LU secretion during the luteal phase is pulsatile. This mode of LU release is similar to that observed in ovariectomized and intact mares during the transition into the breeding season. However, in both ovariectomized and intact mares LU pulses are not readily detected during the breeding season with the exception of the luteal phase in the intact mare. Collectively, our studies suggest that in the mare either tonic LU secretion can occur in both a pulsatile and nonpulsatile mode, or in some physiologic states the collection of blood samples from the jugular vein may not be entirely representative of the release of LU by the anterior pituitary gland. ACKNOWLEDGMENTS We thank the staff of the University Farm for care of the animals, and Drs. T. M. Nett, L. E. Reichert, Jr., H. Papkoff. and L. Edgerton for providing reagents used in the radioimmunoassays. In addition, we thank Ms. S. Barrows for assistance in conducting the radioimmunoassays. Downloaded from on 19 January 018

8 PULSATILE LB RELEASE IN POSTPARTUM MARES 33 REFERENCES Baird, D. T. (1978). Pulsatile secretion of LH and ovarian estradiol during the follicular phase of the sheep estrous cycle. Biol. Reprod. 18: Baird, D. T. and McNeilly, A. 5. (1981). Gonadotrophic control of follicular development and function during the oestrous cycle of the ewe. J. Reprod. Fertil. Suppl. 30: Evans, J. W., Hughes, J. P., Neely, D. P., Stabenfeldt, G. H. and Winger, C. M. (1979). Episodic LB secretion patterns in the mare during the oestrous cycle. J. Reprod. Fertil. Suppl. 7: Fitzgerald, B. P., I Anson, H., Loy, R. G. and Legan, S. J. (1983). Evidence that changes in LH pulse frequency may regulate the seasonal modulation of LB secretion in ovariectomized mares. j. Reprod. Fertil. 69: Gab, R. V. (1981). Pulsatile LH release during periods of low level LH secretion in the rat estrous cycle. Biol. Reprod. : Garcia, M. C., Freedman, L. J. and Ginther, 0. J. (1979). Interaction of seasonal and ovarian factors in the regulation of LH and FSH secretion in the mare. J. Reprod. Fertil. Suppl. 7: Geschwind, I. I., Dewey, R., Hughes, J. P., Evans, J. W. and Stabenfeldt, G. H. (1975). Plasma LH levels in the mare during the oestrous cycle. J. Reprod. Fertil. Suppl. 3:07-1. Ginther, 0. J. (1979). In: Reproductive Biology of the Mare. McNaughton and Gunn, Inc., Ann Arbor, MI, chpt. 7. Ginther, 0. J., Pineda, M. H., Wentworth, B. C. and Nuti, L. (197). Rate of disappearance of exogenous LB from the blood of mares. J. Anim. Sci. 39: Gledhill, B. and Follett, B. K. (1976). Diurnal variation and the episodic release of plasma gonadotrophins in Japanese quail during a photoperiodically induced gonadal cycle. J. Endocrinol. 71: Goodman, R. L. and Karsch, F. J. (1980). Pulsatile secretion of luteinizing hormone: differential suppression by ovarian steroids. Endocrinology 107: Karsch, F. J., Foster, D. L., Bittman, E. L. and Goodman, R. L. (1983). A role for estradiol in enhancing luteinizing hormone pulse frequency during the follicular phase of the estrous cycle of sheep. Endocrinology 113: Knobil, E. (1980). The neuroendocrine control of the menstrual cycle. Recent Prog. Horm. Res. 36: Lincoln, G. A. and Short, R. V. (1980). Seasonal breeding: nature s contraceptive. Recent Prog. Horm. Res. 36: 1-5. Loy, R. G. (1980). Characteristics of postpartum reproduction in mares. Vet. Chin. North Am. : Loy, R. G., Evans, M. J., Pemstein, R. and Taylor, T. B. (198). Effects of injected ovarian steroids on reproductive patterns and performance in postpartum mares. J. Reprod. Fertil. Suppl. 3: Nett, T. M., Holton, D. W. and Estergreen, V. L. (1975). Levels of LH, prolactin and oestrogens in the serum of posr-partum mares. J. Reprod. Fertil. Suppl. 3: Rahe, C. H., Owens, R. E., Fleeger, J. L., Newton, H. J. and Harms, P. G. (1980). Pattern of plasma luteinizing hormone in the cyclic cow: dependence upon the period of the cycle. Endocrinology 107: Whitmore, H. L., Wentworth, B. C. and Ginther, 0. J. (1973). Circulating concentration of luteinizing hormone during the estrous cycle of mares as determined by radioimmunoassay. Am. J. Vet. Res. 3: Yen, S.S.C., Vandenberg, G., Tsai, C. C. and Parker, D. C. (197). Ultradian fluctuations of gonadorropins. In: Biorhythms and human reproduction (M. Fern, F. Galberg, R. M. Richart and R. L. VandeWiele, eds.) Wiley, New York, pp Downloaded from on 19 January 018